#4321: Unexpected stack overflow prevented by superfluous type annotation
---------------------------------+------------------------------------------
    Reporter:  bjpop             |       Owner:              
        Type:  bug               |      Status:  new         
    Priority:  normal            |   Component:  Compiler    
     Version:  6.13              |    Keywords:              
    Testcase:                    |   Blockedby:              
          Os:  Unknown/Multiple  |    Blocking:              
Architecture:  Unknown/Multiple  |     Failure:  None/Unknown
---------------------------------+------------------------------------------
 GHC versions affected:
    6.13.20100915
    6.13.20090922
    6.12.1
    6.10.4

 OS versions tested:
    OS X Snow Leopard
    Ubuntu Linux 9.10 - Karmic Koala

 Description:

 The attached program computes Pi by integrating 4/(1+x*x) between 0 and 1.

 The important line of code is this one:

    h * (sum (map area [1..n]))

 If you compile the program as it is, with this command:

    ghc --make -O2 Pi.hs

 it runs out of stack space when run like so:

    ./Pi 1000000

    Stack space overflow: current size 8388608 bytes.
    Use `+RTS -Ksize -RTS' to increase it.

 For GHC versions 6.10.4, 6.12.1 and 6.13.20090922
 However, if you add what appears to be a superfluous type annotation to
 the line of code, like so:

    h * ((sum (map area [1..n])) :: Double)

 The program runs to completion with no stack overflow:

    ./Pi 1000000
    3.1415926535897643

 Indeed it works just fine for even larger inputs.

 However version 6.13.20100915 overflows the stack regardless of whether
 the type annotation is present or not.

 I don't think the type annotation should make any difference to the
 generated code because GHC should infer that the annotated expression is
 of type Double anyway (if you put an erroneous type in the same place, GHC
 complains that it should be a Double).

-- 
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/4321>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
_______________________________________________
Glasgow-haskell-bugs mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

Reply via email to